{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:BTCZWVEWLACXQY3AEIX5KHM7ZT","short_pith_number":"pith:BTCZWVEW","schema_version":"1.0","canonical_sha256":"0cc59b54965805786360222fd51d9fcceba9bea0c8c4c5bac0c3a17ed3fcd8ea","source":{"kind":"arxiv","id":"2108.12237","version":1},"attestation_state":"computed","paper":{"title":"Evaluating the Robustness of Neural Language Models to Input Perturbations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Matthias Samwald, Milad Moradi","submitted_at":"2021-08-27T12:31:17Z","abstract_excerpt":"High-performance neural language models have obtained state-of-the-art results on a wide range of Natural Language Processing (NLP) tasks. However, results for common benchmark datasets often do not reflect model reliability and robustness when applied to noisy, real-world data. In this study, we design and implement various types of character-level and word-level perturbation methods to simulate realistic scenarios in which input texts may be slightly noisy or different from the data distribution on which NLP systems were trained. Conducting comprehensive experiments on different NLP tasks, w"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2108.12237","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-08-27T12:31:17Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d3b2871c94b5b711fdac41e959204cc66635e90c2e04ee387dec50ea439a8d27","abstract_canon_sha256":"221b53dee72849f2e69749afcec896187ea879c729bcf32ca1686f928b9b8412"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:09:23.586306Z","signature_b64":"mRDZQ2HWr0Jg7Qj30l3PYXuyiLDlj6Arzx/B1CdAxGYmrKfMuJLDKh8/yahWxpGnyK6LDKHes9tisd5G4HFFBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0cc59b54965805786360222fd51d9fcceba9bea0c8c4c5bac0c3a17ed3fcd8ea","last_reissued_at":"2026-07-05T03:09:23.585586Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:09:23.585586Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating the Robustness of Neural Language Models to Input Perturbations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Matthias Samwald, Milad Moradi","submitted_at":"2021-08-27T12:31:17Z","abstract_excerpt":"High-performance neural language models have obtained state-of-the-art results on a wide range of Natural Language Processing (NLP) tasks. However, results for common benchmark datasets often do not reflect model reliability and robustness when applied to noisy, real-world data. In this study, we design and implement various types of character-level and word-level perturbation methods to simulate realistic scenarios in which input texts may be slightly noisy or different from the data distribution on which NLP systems were trained. Conducting comprehensive experiments on different NLP tasks, w"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2108.12237","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2108.12237/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2108.12237","created_at":"2026-07-05T03:09:23.585661+00:00"},{"alias_kind":"arxiv_version","alias_value":"2108.12237v1","created_at":"2026-07-05T03:09:23.585661+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2108.12237","created_at":"2026-07-05T03:09:23.585661+00:00"},{"alias_kind":"pith_short_12","alias_value":"BTCZWVEWLACX","created_at":"2026-07-05T03:09:23.585661+00:00"},{"alias_kind":"pith_short_16","alias_value":"BTCZWVEWLACXQY3A","created_at":"2026-07-05T03:09:23.585661+00:00"},{"alias_kind":"pith_short_8","alias_value":"BTCZWVEW","created_at":"2026-07-05T03:09:23.585661+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2309.10253","citing_title":"GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Prompts","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT","json":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT.json","graph_json":"https://pith.science/api/pith-number/BTCZWVEWLACXQY3AEIX5KHM7ZT/graph.json","events_json":"https://pith.science/api/pith-number/BTCZWVEWLACXQY3AEIX5KHM7ZT/events.json","paper":"https://pith.science/paper/BTCZWVEW"},"agent_actions":{"view_html":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT","download_json":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT.json","view_paper":"https://pith.science/paper/BTCZWVEW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2108.12237&json=true","fetch_graph":"https://pith.science/api/pith-number/BTCZWVEWLACXQY3AEIX5KHM7ZT/graph.json","fetch_events":"https://pith.science/api/pith-number/BTCZWVEWLACXQY3AEIX5KHM7ZT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT/action/storage_attestation","attest_author":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT/action/author_attestation","sign_citation":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT/action/citation_signature","submit_replication":"https://pith.science/pith/BTCZWVEWLACXQY3AEIX5KHM7ZT/action/replication_record"}},"created_at":"2026-07-05T03:09:23.585661+00:00","updated_at":"2026-07-05T03:09:23.585661+00:00"}